Hardware

Two form factors.
One mission.

Choose the hardware that fits your space and scale. Both deliver the same promise: own it once, run unlimited AI forever.

Book a Demo View Specs
Desktop Workstation

ZeroBoxx Standard

Professional AI for teams and offices. Quiet, compact, powerful.

Book a Demo
Compute
GPU NVIDIA RTX PRO 2000 Blackwell Architecture
VRAM 16 GB GDDR7 High-bandwidth memory
CPU NVIDIA Grace 72-core Neoverse V2
System RAM 496 GB LPDDR5X at 6400 MT/s
Storage
Primary 4 TB SSD PCIe Gen4, SED Ready
Connectivity
Ethernet 1 GbE + 10 GbE RJ45
USB 4x USB 3.2 Type-A and Type-C
Software
OS Ubuntu 24.04 LTS Pre-installed
AI Stack NVIDIA AI Dev Tools Ollama, CUDA, cuDNN
API OpenAI-compatible REST API included
Powered by NVIDIA Blackwell
Most Powerful
Rack Server

ZeroBoxx Pro

Data center AI. Maximum inference throughput for demanding workloads.

Book a Demo
Compute
GPU NVIDIA DGX B300 Blackwell Architecture
VRAM 252 GB HBM3e Ultra-high bandwidth
CPU NVIDIA Grace 72-core Neoverse V2
System RAM 496 GB LPDDR5X at 6400 MT/s
Storage
Primary 16 TB SSD 4x 4 TB PCIe Gen4, SED Ready
Connectivity
Network 2x QSFP112 400 Gbps per port
Management BMC / IPMI Out-of-band management
Power
PSU 1600 W Titanium 80 Plus Titanium, C19 inlet
Software
OS Ubuntu 24.04 LTS Pre-installed
AI Stack NVIDIA AI Dev Tools Ollama, CUDA, cuDNN, TensorRT
API OpenAI-compatible REST API, gRPC, Python SDK
Powered by NVIDIA Blackwell

Included with every ZeroBoxx

🔒 Zero data sent to cloud. Ever.
♾️ Unlimited inferences. No rate limits.
🧩 OpenAI-compatible REST API
🤖 Run any open-source LLM model
🛠 Full fine-tuning on your data
📦 Ships pre-configured, ready on day one
🔄 Swap models without restrictions
🌐 LAN deployment, no internet required

Frequently asked questions

What is the difference between ZeroBoxx Standard and ZeroBoxx Pro?

ZeroBoxx Standard is a desktop workstation designed for teams and offices, featuring an NVIDIA RTX PRO 2000 GPU with 16 GB GDDR7 VRAM. ZeroBoxx Pro is a rack-mount server for enterprise data centers, featuring the NVIDIA DGX B300 with 252 GB HBM3e for maximum throughput.

How long does setup take?

ZeroBoxx ships pre-configured with NVIDIA AI development tools and Ollama. Most customers are running their first inference within a few hours of delivery.

Can I fine-tune models on ZeroBoxx?

Yes. Both ZeroBoxx Standard and Pro support full model fine-tuning using your own data. NVIDIA AI Dev Tools including the NeMo framework are pre-installed.

Ready to own your AI?

Book a demo to see ZeroBoxx in action and get a configuration recommendation for your team.

Book a Demo